The quality of care in U.S. hospitals continues to alarm. As a case in point, medical errors persist at a distressing rate and incur enormous costs, despite extensive publicity and considerable research. Because healthcare is a complex sociotechnical system, these medical errors generally have no simple, single cause. Instead, the errors typically are the result of system failures that involve many individuals, many contributing factors, and encompass multiple levels of the health care system. In our previous research we have shown that computational modeling using OrgAhead is a valuable research tool for dealing with these types of complex, multi-level problems. Building on that work, we propose using a network-based theory of organizations as complex adaptive systems to develop desktop decision support tools and methods that help managers (a) assess their unit's current performance, and (b) predict the impact of potential unit innovations to improve targeted safety and quality outcomes. The purpose of the proposed project is to develop a simple virtual environment that allows nurse managers to assess, through a dynamic network analysis decision support tool (DyNADS), the organizational health of their patient care units and then engage in strategic planning by using automated "what-if" analysis techniques to test, on the virtual units, the likelihood of potential innovations to improve their actual units'safety and quality outcomes. Specific aims are to: 1. Refine and streamline data collection requirements for DyNADS to fit within the unit management workflow. 2. Design a prototype desktop dynamic network analysis decision support tool (DyNADS) that facilitates: (a) identifying structures or communication patterns that are the basis for system failures in acute care settings and (b) testing the likelihood of potential innovations to improve safety and quality outcomes. 3. Evaluate DyNADS'validity, scalability, and usefulness through a field test. 4. Evaluate the feasibility of implementing DyNADS or nursing units.